One very brief comment on yesterday's BBC story about racism in the labour market. As usual, the BBC did not elaborate in their story on exactly how the data were obtained, and the verbal description of their results is unsatisfying:

CVs from six fictitious candidates -- who were given traditionally white, black African or Muslim names -- were sent to 50 firms by Radio Five Live.

[...]

Many of the firms were well known and the jobs covered a range of fields, Radio Five Live said.

All the applicants were given the same standard of qualifications and experience, but their CVs were presented differently.

Almost a quarter of applications by two candidates given traditionally ``white'' names - Jenny Hughes and John Andrews - resulted in interview offers.

But only 9% of the ``Muslim'' applications, by the fictitious Fatima Khan and Nasser Hanif, prompted a similar response.

Letters from the ``black'' candidates, Abu Olasemi and Yinka Olatunde, had a 13% success rate.

(Having read that, you might -- for instance -- be wondering how, out of 50 cases, they can have obtained a particular result in an odd-numbered percentage of trials; or, more generally, whether these results really tell us anything.)

(I mentioned some of this in comments on Matthew's web log; the discussion there is mostly unenlightening, but it's worth reading Tom's contribution, since he actually knows something this area, which the other participants mostly, I think, do not.)

In fact, the BBC have presented the raw data from the study along with examples of the CVs they sent to the various employers. The aggregate results were:

implied race of applicant	number of applications resulting in invitation for interview	not invited for interview	no reply received	application `lost'
white	23	22	55	0
asian	8	20	66	6
black	13	22	65	0

At the simplest level of analysis, these results provide strong evidence (chi² = 21.2731, p = 0.0011) for the hypothesis that the result of a job application is influenced by the apparent race of the candidate. Race doesn't significantly influence whether a rejected candidate is told to piss off (`not invited') or simply never contacted again, but it is significant (p = 0.01681) in determining whether their application is `lost'.

It is of course true that a test of statistical significance does not of itself show that the conclusions of the study are accurate: the design of the study must also be sensible. But there's nothing obviously wrong with the methodology here, and as Tom points out the conclusions of this study are similar to those of various previous studies (sadly those he refers to -- Half a Chance, Still? published by the Nottingham and District Racial Equality Council in 1994; We Regret to Inform You... published by the CRE in 1996; and Racial discrimination against doctors from ethnic minorities, BMJ 306:691-2, 1993 -- are not available on-line).

Anyway, it turns out that, British employers are racist (at least where it comes to black african and Asian job applicants). This sucks pretty badly, and I'd say the fact that there exist purportedly intelligent apologists (they know who they are) for this state of affairs sucks almost as much.

(On a tangentially related note, this story in yesterday's Guardian describes how hard it is to get a booking in a restaurant if your surname is `bin Ladin', and also tells us that the enemies of civilisation are terrible at interior decor:

[Carmen bin Ladin, estranged wife of Osama's brother, Yeslam bin Laden] cattily remarks in the book that the Bin Ladens have no taste in interior decor, all gold taps and terrible paintings....

In other news, don't take up photography if you're a US citizen with dark skin. Elsewhere, John Band elaborates on one of my darker fears, illustrating it with a short visit to the edge of reason, personified by crank right-wing web loggers. Actually I disagree with John on one point; he writes,

But we also need to ensure that in the horrible event that the intelligence fails [and there is a further serious terrorist attack], we don't let the macho headcases dig our graves for us.

-- my fear is not that we will have our graves dug for us, but that if the West overreacts and embarks on a war of genocide, it will be brutally effective in prosecuting it.)